Skip to main content
Fig. 3 | BMC Bioinformatics

Fig. 3

From: Topology preserving stratification of tissue neoplasticity using Deep Neural Maps and microRNA signatures

Fig. 3

SOM lattice showing mapping of all a neoplastic tissue, b non-neoplastic tissue, and c neoplasticity status and tissue-of-origin from nine selected tissues-of-origin for both training (top) and test (bottom) data. The x and y axes represent SOM coordinates. The color of the heatmaps indicates the density of samples in columns a, b. In column c, circles and squares represent non-neoplastic and neoplastic samples, respectively, with the size of the marker proportional to the number of samples of the classes mapped to that location. From columns a, b, the DNM is able to stratify neoplastic and non-neoplastic tissue through mapping to different areas of the SOM. Clear cluster-centers can be seen in both of these images in red. Column c shows clear clusters of similar tissue and neoplasticity status. Models generalize well to test data as seen from the top and bottom rows of the figure. In the legend, * indicates that these neoplastic tissue types consist of some/all neuroendocrine tumour samples

Back to article page